MARINE ECOLOGY PROGRESS SERIES Published July 15 Mar. Ecol. Prog. Ser. l

Experimental evaluation of commercial-scale enhancement of Haliotisiris populations in New Zealand

David R. Schiel

Department of Zoology, University of Canterbury, Private Bag 4800, Christchurch 1, New Zealand

ABSTRACT: The viability of enhancing populations of abalone iris by placing hatchery-raised juveniles into natural habitats was evaluated at the Chatham Islands, New Zealand. Eighty thousand abalone, from 3 to 30 mm (shell length) in size, were seeded into 8 sites during 1990 and 1991 and sampled for survival and growth at times ranging from 5 to 23 mo. Annual mortality rates were 27.6 to 98.8 'X and instantaneous mortality ranged from 0.323 to 4.440. Mortality was greatest in sites affected by sand movement, which partially buried juvenile habitat. Desirable characteristics of sites for en- hancement were identified so that the risk of mortality due to habitat shifts can be greatly reduced in the future. Average annual growth increments ranged from 13.5 to 23.8 mm, but the lower value may be conservative because of the varying growth periods that seed were in place at the different sites. Preliminary economic analyses showed that, based on the current price of abalone, 3 individual sites should produce positive financial returns on the future harvest of abalone at the minimum legal size of 125 mm. Although 5 sites will have negative financial returns, the overall return of the 8 sites should exceed the original investment.

INTRODUCTION One area where marine restoration may have a chance of success is in invertebrate fisheries, which are There is considerable interest in the methods and susceptible to overfishing (Harrison 1986). This is par- means of restoring marine populations that have been ticularly true of the commercially valuable abalone damaged or depleted in some way (Thayer 1992). To fisheries, which have suffered serial depletion in many be successful, this type of work requires a considerable countries, including New Zealand (Shepherd et al. knowledge about the life history of target organisms 1992b). The possibility of augmenting heavily ex- and an understanding of their ecology. In particular, it ploited abalone populations to help ensure their long- is necessary to know the links between early life stages term sustainability by enhancement with hatchery- and adult stages in terms of habitat requirements, raised seed has been explored in a few countries (Saito growth and survival. These links are far from clear for 1984, Tegner & Butler 1989). The interest in establish- most . Especially in long-lived species, the rela- ing such a programme in New Zealand is more recent tionship between the numbers of juveniles and adults and has been in response to the perception that natural may break down under the influences of resource recruitment is irregular in many fishing sites limitation, predation, and environmental conditions (Sainsbury 1982a) and that the current high levels of (Underwood & Denley 1984, Connell 1985, Keough fishing effort in the comn~ercialand recreational fish- 1988, Jones 1990). These relationships have consider- ery may not be sustainable (Schiel 1992b). Population able importance in cornmercially-exploited species for enhancement, in conjunction with various manage- which economic value is attached to management ment initiatives, was seen as a means of significantly decisions. Moreover, any attempts at restoration or increasing recruitment to the fishery. enhancement of fished populations must be econorm- Attempts at enhancing abalone populations have cally viable. produced variable results (reviewed in Tegner & Butler Mar. Ecol. Prog. Ser. 97: 167-181,1993

1989). In southern California (USA),it appears that few generally occur in water depths <5 m (Poore seeded juveniles of HaLiotis rufescens survived to har- 1972c, Sainsbury 1982a). Sites where extensive sand vestable size despite many releases (Tegner & Butler and rock movement occur are usually poor habitat for 1985, 1989). Most seeded abalone were never seen juveniles. Rocky reefs that provide good juvenile habi- after release. Significant mortality was due to preda- tat are relatively common around the Chatham Islands, tion by crabs, lobsters and octopus, to the stressed con- New Zealand, where there is a substantial commercial dition of the hatchery-reared juveniles after their paua fishery (Schiel 1992b). This study was designed transport to seeding sites, and to alterations of habitat as a large-scale experiment in which growth and sur- following storms. Trials in Japan with H. discus hannai vival of seeded paua were assessed within a range of have also had variable results but successes have been sites around the island. The objectives were to deter- reported, with some seeded abalone eventually being mine the habitat characteristics suitable to enhance- captured in the commercial fishery (Saito 1984). Where ment on a commercial scale and to provide information recapture rates of seeded abalone were high, the size on whether seeding is an economically viable form of of seed at release was generally 230 mm (Saito 1984, paua population enhancement. Tegner & Butler 1989). New Zealand native abalone (or paua) Haliotis ins are relatively slow-growing, with average increments MATERIALS AND METHODS in shell length of 20 to 25 mm yr-' as juveniles (Poore 1972~).In a hatchery, the cost of raislng smaii abaiorie Study sites. The Chatham Islands are situated is considerably less than raising larger juveniles 700 km from mainland New Zealand at 44'S, 176'W because of the feed, attention, and time required to on the eastern edge of the Chatham Rise (Fig. l), a sub- maintain cultures of larger . Therefore, much marine ridge 250 to 500 m deep. The Chathams consist greater numbers oi smaiier individuals can be cultured of 2 main islands and several rocky outcrops. They for a given cost. Because it is necessary for seeding comprise a major fishery management area for Haliotis projects to be financially viable, the trade-offs between iris, with about 292 t caught annually (Schiel 1992b). costs of seed production, the size at release, and the Paua of all sizes are very abundant around the islands survival to harvestable size must be considered (Schiel on virtually all rocky inshore reefs (Schiel unpubl. 1989). In New Zealand, there is only one small hatch- data). This study was done on the main island, ery capable of raising juvenile paua. Growing them to Chatham Island. larger sizes is hampered by the lack of a cheap and I searched extensively for good juvenile habitat, abundant food. Because of the limitations on space and seeking sites of at least several hundred m2 with small food, any programme seeking to use large numbers of boulders on rocky reefs. This type of large-scale juve- hatchery-raised paua for enhancement of populations nile habitat is patchy in occurrence around Chatham is constrained to using small juveniles (<20 mm), which can be raised in the available facility (Tong & Moss 1992). Using seed paua of 8 to 15 mm in size pro- vides a reasonable balance between costs and returns (Schiel 1992a). In small-scale experiments, the recov- ery of 20 mm paua was slightly greater than 12 mm CHATHAM ISLAND paua (Schiel unpubl. data) but the cost of producing the smaller paua was considerably less. -50 S As for many abalone species, Haliotis iris juveniles (

Table l. Descnpt~ons and characteristics of enhancement Port Hutt and Whangamoe sites are more protected by sites at Chatham Island. Wind direction is that of greatest headlands. However, during southerly storms the full exposure at each site force of waves several meters in height can strike these sites. Kaingaroa and Owenga are affected by winds Site Exposure Characteristics from north through southeast. Hatchery methods. Eighty thousand juvenile paua Port Hutt (Alabama) S-SW Extensive boulder and small rock (10000 per site) were required for the project, with Owenga N-E Some reef, smaller the intention of seeding them over a 6 mo period. rock, large sand patches However, the production schedule of seed paua was Port Hutt (2) S-SW Extensive reef, some not met and seeding was therefore extended over small rock 1990-1991 (Table 2). Several batches of paua juveniles Whangamoe S-SW Some reef, large were grown during this period at the Ministry of boulders and rock Agriculture & Fisheries hatchery in Wellington. Waikaripi NW-SW Steep reef, some smaller rock Because of the geographic isolation of the Chatham Karen Wreck NW-SW Some reef, boulders and Islands and the possibdity of a separate genetic stock smaller rock there, adult paua from the Chathams were used for Pumphouse NW-SW Some reef, boulders and spawning in the hatchery. At least 8 males and 8 smaller rock females were used in each spawning to minimise the Kaingaroa W, N-E Reef, some smaller rocks, risks of loss of genetic variation (Smith & Conroy 1992). extensive sand patches The techniques used to spawn adults, rear and settle larvae, and raise juvenile paua have been described elsewhere (Tong & Moss 1992). It took 4 to 6 mo to Island and many areas were excluded from the seed- raise paua to an average size of ca 10 mm. Because of ing experiment because of severe exposure, inade- hitations on available rearing tanks, however, paua quate access, or inappropriate habitats. Much of the in- were settled and grown in separate batches. The first shore habitat of Chatham Island is composed of sand, lot of 10000 was raised during 1989 and seeded in especially in the long expanses of Hansen Bay, Petre January 1990 (Table 2); 30000 were available in May Bay, and the north coast. Other areas had difficult ac- 1990 and seeded at 3 sites; 30000 were grown during cess (the northwest coast) or were continually exposed late 1990 and seeded in February 1991; 10000 were to severe sea conditions (the south coast). Conse- grown in late 1990 and seeded in May 1991. quently, study sites could not be spaced evenly around Although the intention was to use seed paua of a nar- the island. Eight were chosen that represented a range row size range (8 to 12 mm), this proved to be difficult of conditions, but all had rocky reefs with extensive to achieve for several reasons: (1) paua were raised in juvenile habitat (Fig. 1, Table 1).It was anticipated that different tanks and during different times in the hatch- some sites would prove to be inappropriate for large- ery; (2) sometimes it was necessary to hold abalone for scale seeding. In particular, Kaingaroa and Owenga a few months before field conditions were suitable for were surrounded by sand, which can shift during seeding; (3) there was considerable variability in the storms. growth of paua both within and between rearing Chatham Island is extremely exposed to wind and tanks. For example, paua from 2 settlements 10 d apart sea conditions. Prevailing southwesterly winds pro- during April 1990 and grown in 5 different tanks were duce a swell that refracts around Point Durham. The significantly different in size in the hatchery in

Table 2. Haliotis iris. Summary of seeding history of abalone at Chatham Island sites

Site Depth Seeding date Final sample Area Number (m1 seeded (m2) seeded

Port Hutt (Alabama) 2-3 10 Jan 90 12 Dec 91 507 10 000 Owenga 1-3 11 May 90 23 May 91 280 9 900 Port Hutt (2) 2-3 12 May 90 9 Dec 91 150 9 850 Whangamoe 2-3 13 May 90 12 Dec 91 340 9 850 Waikaripi 3-4 20 Feb 91 11 Dec 91 800 7 900 Karen Wreck 1-3 22 Feb 91 11 Dec 91 900 10018 Pumphouse 1-3 22 Feb 91 11 Dec 91 500 8 523 Kaingaroa 1-3 27 May 91 10 Dec 91 625 10 000 Mar.Ecol. Prog. Ser. 97: 167-181, 1.993

February 1991 (ANOVA: F4,1101= 137.46, p i 0.001). For transferral to study sites, seed paua were re- The average sizes of paua in the 5 tanks ranged from moved from live wells, counted, and placed back into 11.3 to 18.7 mm. I attempted to lessen any potential shipping tubes which were packed with sealed bags of effects of variable seed size by mixing all of the paua ice into cartons. When on site, about 1000 paua at a available for a seeding period before placing them into time were put into mesh bags carried by 2 divers, who habitats. placed small groups of juveniles by hand beneath When ready for seeding, paua were removed from rocks throughout each seeding site. Because small the rearing tanks by hand, using a stream of warm paua usually cling together in clumps when trans- water to loosen abalone from the surface. Paua were ported, divers ensured that individuals were separated then packed into shipping tubes with wet seaweed and attached to rocks. The process of transferring (Gracdaria sordida) (methods and tubes described in l0 000 seed paua took about 4 man-hours of underwa- Tong & Moss 1992). Up to 1000 paua were placed into ter work at each site. The target seeding density was each tube. The tubes were held in the hatchery ca 50 to 70 m-' of juvenile habitat, although total den- overnight in a seawater tank that drained itself every sities were lower when the entire site was taken into 2 h. On the morning of the flight to the Chatham account. The target density was comparable to that at Islands, tubes were packed into styrofoam cartons with which larger juveniles (>20 mm) occur in good habitat sealed bags of ice for chilling. At the Chathams, they at the Chatham Islands and mainland New Zealand were taken to Waitangi and opened in a fish packing (Schiel unpubl. data). In most cases, paua readily at- shed. The tubes were piaced into iive weiis used for tached to rocks and appeared to be in qood condition. keeping lobsters alive before export. These wells had a Post-seeding surveys. Paua were left for variable large capacity flow-through water system. The time times before post-seeding surveys because trips to the from hatchery to live wells was about 7 h. Chatham Islands had to be timed around periods when Most paua arrivecl,in good cor;ditior,, but there was new seed was available and weather conditions were some mortality. In one shipment where all paua were suitable for diving. Because of the cryptic nature of counted there was 1.5% mortality (167 dead out of small paua, the undersides of rocks were searched dur- l0 962). In another shipment, 4 tubes that were packed ing surveys. Forty randomly placed 0.25 m2 quadrats at a density of 1000 paua per tube suffered 47 O/O mor- were sampled on each survey. Hatchery paua were tality. This high mortality was an exception and paua identified and measured in situ.Searches were done in were afterwards packed at densities no greater than nearby areas of juvenile habitat to determine if seed 500 per tube. A greater problem was the estimation of paua had spread from the immediate seeding site. numbers of paua in each shipment. Paua were not in- Small paua were easy to identify as hatchery-reared dividually counted into tubes by hatchery staff before because they had a characteristic bright green mark on shipment. Instead, 5 randomly selected groups of 500 the apex of the shell and they usually produced a pro- juveniles were counted and weighed, and the mean nounced growth after placement into natural habi- weight was used to estimate batches of 500 for each tats. These marks became more difficult to discern, tube. When opened at the Chatham Islands it was clear however, after about 2 yr, or when paua reached more that some estimates were too high. Because of these than ca 50 mm in length. Paua whose origin was in inaccuracies, all paua were individually counted after doubt were recorded separately and it was subse- shipment to the Chatham Islands from May 1990 quently decided to assign 50% of these as being of onwards. hatchery origin. This percentage was arbitrarily Seeding methods. The 8 study sites were sampled chosen but is probably conservative in estimating for paua before seeding. Forty 0.25 m2 quadrats were numbers of hatchery paua because only paua that had randomly placed in each site and all paua were hatchery-type markings were included. Paua of doubt- counted and measured. Juvenile paua were present at ful origin usually comprised only a few individuals. most sites but they were not usually of similar sizes to Growth of seed paua was recorded as size frequen- those that were seeded (see 'Results'). I investigated cies through time. Proportional survival of the entire the possibility of establishing a control site near each seed cohort at a site was more difficult to determine. seeding site to serve as an unseeded comparison in the Sampling for juveniles can be quite destructive of event that hatchery paua were not subsequently iden- under-boulder habitats and it was not feasible or desir- tifiable after seeding These unseeded sites were even- able to overturn entire seeding sites to collect all seed tually omitted because hatchery juveniles were distin- paua that may have been present. Subsampling pre- guishable from natural paua, and the unseeded sites sented problems in that paua are very patchy in occur- were invariably different from the corresponding rence and therefore there is a high variance in any seeded ones in both habitat structure and paua sizes, quadrat sampling of numbers. To estimate survival, the thereby obviating their use as 'controls'. number of seeded paua were averaged over the total Schiel: Enhancement of abalone populations 171

area of a site to give a benchmark for the PORT HUT(ALABAMA) seeding density (see Tables 2 & 3). Sam- A. Pre-seeding sizestructure pling means through time were compared to this seeding density and the 95 % confi- l0 l dence intervals were computed using the standard error of samples. Economic analysis. A preliminary eco- nomic analysis was done to determine whether there is any likelihood of seeding being economically viable at the Chatham 0 50 100 150 Islands. The assumptions of the model are Shell length (mm) included in the 'Results'. To estimate future B. Initial seed sizes (l0 Jan 90) C. At five months (5 May 90) numbers, the mortality rates recorded for seed paua at each site were used for the ju- venile period (3 yr) and an annual mortality rate of 20% was used for the next 3 yr (Sainsbury 1982a). The future value of the money used to purchase seed paua was used as the standard against which future profits or losses were gauged. The internal rate of return (IRR)of the final financial re- Shell length (mm) Shell length (mm) turns against the original investment was D. At one year (23 Feb 91) E. At two years (12 Dec 91) 30 assessed. 25 25 ! RESULTS

Because sites were seeded at 4 times z l~ ,

(Table 2), results will be discussed within 5 l each seeding episode. o I-- 0 10 20 30 40 50 60 70 0 10 20 30 40 50 60 70 Shell length (mm) Shell length (mm) Site at 2 years F. Growth G.Survival

A site at Port Hutt, locally called M I Alabama, was seeded in January 1990 and $ 40 sampled 3 times during the 23 mo to E ! December 1991. The pre-seeding survey showed that juveniles were present, with g 1

l0 1 7.5 "h of them being <15 mm in shell length V) (Fig. 2A). The density of all paua at the site G 0 8 01 , IOJAW5&90 23~:bb91 ?~DZZ 10~dns0~U&BO ZJF&Q~ 120dc91 was 10.0 ind. m-' (SD = 9.72). Date Date Seed paua ranged from 3 to 13 mm shell length (mean = 8,6 SE = 0.08; Fig. 2B). Fig. 2. Hahot~siris. Size-frequencies of abalone at Port Hutt (Alabama) (A) There was approximately 507 m2 of habitat prior to seeding and (B to E) during the seeding experiment. (F) Average into which were placed 10 000 paua, giving a shell length (f 1 SE) of seeded paua through time. (G) Percentage survival seeding density of 19.7 ind. m-' (Table 2). of the original cohort (+ 95 Oh confidence interval) through time After 5 mo these grew to an average size of 17.7 mm (SE = 0.55; Fig. 2C). The range of sizes re- cernible green marks on the shell which are character- flected that at seeding. However, there was consider- istic of seed paua. Furthermore, there was no evidence able variation in sizes at 1 yr (Fig. 2D). Paua tend to grow of recent recruits during the May 1990 sampling period. rapidly during the warmer months (Poore 1972~)and it By December 1991, seed paua ranged in size from is therefore likely that paua at this site had undergone 2 23 to 63 mm (Fig. 2E). There was almost a bimodal dis- periods of growth. Many paua had barely grown and tribution in sizes, with peaks at ca 27 and 45 mm. Many some were still S20 mm, while others were as large as of the seed paua were clearly growing at a slow rate. 48 mm. It is unlikely that the smaller paua were the Overall, seed paua grew from an average of 8.6 mm to result of natural recruitment. They had clearly dis- 42.8 mm in 23 mo (Fig. 2F). 172 Mar. Ecol. Prog. Ser. 97: 167-181, 1993

Estimates of percentage survival showed apparently Paua were patchy in occurrence on the final sam- anomalous results (Fig. 2G). Approximately 39 % pling date and there was a large confidence interval (f 13.7 %) of the original 10000 seed paua survived by for the survival estimate. Nevertheless, survival was May 1990, ca 19 % (f 6.9 %) at 1 yr, and then an average unlikely to be lower than 25 % (the lower confidence of 54 % (f 29.1 %) at 23 mo. This apparent drop and rise interval, CI) after almost 2 yr. in survival was probably the result of the non-destruc- tive sampling procedure. Because paua become less Sites at 1.5 years cryptic and move into more accessible parts of the habi- tat as they grow, they are more likely to be encountered Three sites were seeded during May 1990 (Table 2), as they get larger. Many of the larger rocks in seeding two of which could be sampled only once, 19 mo later sites were situated over smaller rocks and stones with in December 1991. The Owenga site is on the eastern many interstices that may conceal small paua. Larger ju- side of Chatham Island where there is predominantly veniles tend to move onto the undersides of larger rocks sand habitat. This site has low-lying rocks and boul- and therefore can be seen more easily. The doubtful ori- ders interspersed around a sandy lagoon. An outer gin of some paua >50 mm could also have contributed to rocky reef protects this site from the full force of east- the higher survival estimate on the final sampling date. erly storms. The pre-seeding survey showed that there This effect should be slight, however, because only 9 % were some juvenile paua and many smaller adults in of the recovered paua were of doubtful origin. the area (Fig. 3A), with an overall mean of 3.4 ind. m-2 (SD = 5.52). Tie habitai drea available to OWENGA juvenile paua was ca 280 m2 (Table 2) into which 9900 paua were placed, giving a seed- A. Pre-seeding size structure ing density of 35.4 ind. m-2. I The seed paua ranged in sizs from 3 to 22 mm with a mean of 12.2 mm (SE = 0.95; Fig. 3B). The bimodal nature of the size fre- quencies was due to the paua being grown in different tanks for different periods of time (see 'Methods'). They were sampled almost exactly 1 yr later in May 1991. By then their sizes ranged from 17 to 44 mm (Fig. 3C). The Shell length (mm) average shell length was 34 mm (SE = 1.36; B. Initial seed sizes (l l May 90) C. At one year (23 May 91) Fig. 3D). This was similar to the sizes attained 80 1 10 1 after 1 yr at the Port Hutt (Alabama) site, although the average initial size was larger. Survival after 1 yr was estimated to be 9.9% (f 4.44; Fig. 3E). There had clearly been some sand movement at this site and many of the low-lylng rocks had become compacted into the substratum, rendering 0 5 10 15 20 25 30 35 40 45 50 0 5 10 15 20 25 30 35 40 45 M them no longer suitable as juvenile habitat. Shell length (mm) Shell length (mm) The second site at Port Hutt had ca 150 m2 . D.Growth - . E. Survival of juvenile habitat. It was seeded with 9850 ci 1 paua, giving an overall seeding density of 65.7 ind. m-2 (Table 2). The pre-seeding size structure showed that some juvenile paua were naturally present at this site, with 5 % 5 p 15 of these 520 mm in length (Fig. 4A). The 2- l0 overall pre-seeding density of paua was 05 2.7 ind. m-2 (SD = 5.24). 0 11 May90 23 May 91 11MayW 23 May 91 The sizes of seed ranged from 5 to 22 mm, Date Date with an average shell length of 12.2 mm (SE = 0.95; Fig. 4B), similar to the distribution of Fig. 3. HaLiotis iris.Size-frequencies of abalone at Owenga (A) prior to seed sizes at Owenga. There was only 1 post- seeding and (B, C) during the seeding experiment. (D) Average shell length (* 1 SE) of seeded paua through time. (E) Percentage survival of seeding survey, 19 mo later in December 1991 the original cohort (f 95 % confidence interval) through time (Fig. 4C). Sizes then ranged from 21 to 53 mm. Schiel: Enhancement of abalone populations 173

The bimodal nature of sizes was still evident. PORT HUlT (2) Paua were on average 34.1 mm in shell length (SE = 1.24; Fig. 4D), the same size as A. Pre-seed~ngsize structure at the Owenga site after 1 yr. Survival at Port Hutt (2) appeared to be poor (mean = 7.6 %, _+ 3.78; Fig. 4E). The 6 juvenile habitat at this site, however, was very complex. It was surrounded by large 2 4 boulders and reef, and was beneath a canopy of fucalean algae. Furthermore, it 0. was contiguous with other patches of juve- o .X 40 €0 80 100 120 140 160 Shell length (mm) nile habitat and it is probable that seed paua migrated from the immediate site. B. Initial seed sizes (12 May 90) C. At 1.5 years (9 Dec 91) Some seed paua were found up to 50 m 80 ~liL from the seeding position. M) Whangamoe was the third site seeded n& n during May 1990. It had a juvenile habitat 540 ;: area of ca 340 m* into which 9850 paua z were seeded (average density = 29 ind. m-2; 20 5

Table 2).The pre-seeding survey showed an 0 0 overall density of 3.1 paua ind. m-' (SD = o 5 10 15 20 25 30 35 40 45 50 o 5 10 15 20 25 30 35 40 45 50 55 60 4.56), with few juveniles present (Fig. 5A). Shell length (mm) Shell length (mm)

Seed size ranged from 5 to 22 mm (mean = D. Growth E. Su~kal 11.8, SE = 0.90; Fig. 5B), similar to the other 2 45 . UJ 40 sites seeded in May 1990. A survey 1 yr later $ showed an average size of 36.6 mm (SE = ,30 -g 0.63), with paua ranging from 20 to 48 mm in g 25 . =m shell length (Fig. 5C). Sizes were more van- 20 gC 15 . .2 40 able at 19 mo (Fig. 5D) with paua ranging 2 lo \ - / from 22 to 64 mm in size. Most growth 5. 2 occurred during the first 12 mo, with only a v, 0 l__-- -- 8 0 12May 90 9k91 12 May 90 9k91 small change in average size during the win- Date Date ter and spring of the final 7 mo (Fig. 5E). Survival at 1 yr was estimated to be 18.2 % Fig. 4. Haliotis iris. Size-frequencies of abalone at Port Hutt (2) (A) prior to (+ 6.92) and at 19 mo was 24.2 % (f 9.18) seeding and (B, C) during the seeding experiment. (D) Average shell (~i~.SF). once again, where more than 1 length (+l SE) of seeded paua through time. (E) Percentage survival of the original cohort (f 95% confidence interval) through time post-seeding sample was obtained, the great- est decline in recovery occurred between seeding and the first subsequent survey. (Fig. 6C). The average size of seed was 15.5 mm (SE = 0.28), which grew to an average of 27.0 mm (SE = 2.10) Sites at 10 months 10 mo later (Fig. 6D). Percentage survival was estimated to be 19.2% (k 10.35) by December 1991 Three sites were seeded in February 1991 and (Fig. 6E). checked again in December 1991 (Table 2). AU, of The Karen Wreck site had an initial density of 14.3 these sites were along the coast northeast of Point paua m-* (SD = 12.29). These were mostly adult paua Durham. but 6.5 % of them were $30 mm in size (Fig. ?A). The Waikaripi had an initial density of 6.8 paua m-* (SD habitat area was ca 900 m', into which were placed = 3.64). A few juveniles were present, with 6.7 % of 10018 hatchery paua, yielding a seeding density of them 130 mm in length (Fig. 6A). There was a juvenile 11.1 ind. m-2. Again, there was a large variation in the habitat area of ca 800 m2 into which 3900 paua were sizes of seed, which ranged from 7 to 29 mm (Fig. ?B). placed, giving an overall seeding density of 9.9 ind. The post-seeding survey in December 1991 showed a m-2. There was considerable variation in the sizes of size structure similar to Waikaripi, with sizes ranging seed, ranging from 6 to 31 mm (Fig. 6B). This variation from 18 to 43 mm (Fig. ?C). The initial size of seed was maintained as paua grew during the following averaged 14.8 mm (SE = 0.56), which grew to an aver- 10 mo, when paua ranged from 18 to 49 mm in size age of 25.6 mm (SE = 0.88) 10 mo later (Fig. ?D). Mar. Ecol. Prog. Ser. 97: 167-181, 1993

Survival at this site was good after 10 mo. The esti- Site at 7 months mated survival was 46.7 %. although the confidence interval was large (f 20.09; Fig. ?E). The last lot of 10 000 paua were seeded at Kaingaroa The Pumphouse site was the last place seeded during May 1991. The juvenile habitat area was esti- during February 1991. It had an initial density of mated to be 625 m', yielding an overall seeding density 5.0 paua m-2 (SD = 6.04). Many juvenile paua naturally of 16 ind. m-2 (Table 2). The pre-seeding survey occurred in this site, with almost 50 % of them being showed an initial density of 1.2 paua m-' (SD = 3.76). 530 mm in size (Fig. 8A). The site had a juvenile Few juveniles were present and none were in the habitat area of ca 500 m2 into which were seeded smaller size categories (Fig. 9A). The seed size ranged 8523 paua (seeding density = 17 per m'). The initial from 6 to 23 mm, with most paua at 9 to 12 mm seed size had a wide variation, ranging from 9 to (Fig. 9B). Seven months later these ranged in size from 31 mm, which was similar to that at the other 2 sites 16 to 26 mm (Fig. 9C). The average initial seed size in (Fig. 8B). These grew to between 19 and 53 mm by May was 11.6 mm (SE = 1.16). Paua grew to an average December 1991 (Fig. 8C). The average size of seed of 20.7 mm (SE = 1.00) by December (Fig. 9D). Survival paua in February 1991 was 17.1 (SE = 0.40), which at this site was poor (Fig. 9E). There was considerable grew to an average size of 31.3 mm (SE = 1.47) by sand movement over the site by December and many December (Fig. 8D). Survival was estimated to be of the rocks in the seeding area had been covered or 29.2 % (f 10.98) 10 mo after seeding (Fig. 8E). partially buried by sand. Survival was estimated to be 7.5% jf 7.35). Because of the abundance WHANGAMOE and movement of sand around this site, it seems unlikely that many of the seed paua A. Pre-seeding size structure B. Initial sizes (13 May 90) will survive. 6. 701

Overall patterns

Comparisons of survival and growth among seeding sites can be made only ten- tatively because of the different ranges of seeding sizes, the varying times of seeding, and the varying periods between seeding Shell length(mm) Shell length (mm) and the final survey. The annual mortality C. At one year(28 May91) D. At 1.S years (1 9 Dec 91) rate and the annual instantaneous rate of

30 1 l4 l mortality were calculated for each site (Table 3). Annual mortality ranged from 27.6 to 98.8 '?h and instantaneous mortality ranged from 0.323 to 4.440. The major note- worthy feature is that the site where paua had been in place for the longest time, Port Hutt (Alabama), also had the lowest mortal- 0 5 10 15 X) 25 30 35 40 45 50 0 5 10 152025303540455055606570 ity rates. This translates to an average Shell'length(mm) Shell length(mm) annual survival of 72 % over the f~rst2 yr. F. Survival The greatest mortality may occur in the first few months after seeding (cf. Fig. 2G), al- though an, accurate assessment of mortality during the first year can be confounded by the cryptic nature of small paua. It is possi- ble that future sampling will show greater survival at several sites, particularly Port Hutt (2), Waikaripi, Karen Wreck and

13 May 90 28 May 91 19Dec g1 13May90 28My91 1QDac91 Pumphouse. Date Date Annualized average growth rates varied between 13.5 to 23.8 mm yr-' (Table 3). Fig. 5. Haliotisins. Size-frequencies of abalone at Whangamoe (A]pnor to These may be biased because the warmer seeding and (B to D) during the seeding experiment. (E) Average shell length (f l SE) of seeded paua through time. (F) Percentage sunrival of months, when growth rates are greatest, are the original cohort (f Y5"i confidence interval) through time not equally represented among the sites. Schiel Enhancement of abalone populations 175

For example, growth rates appeared to be WAIKARIPI greatest at 3 [Port Hutt (Alabama), Owenga, Whangamoe] of the 4 sites in which average ,,A. Pre-seed~ngsize structure growth increments could be calculated for a l 12 mo period, rather than having to extrapo- late growth to cover a year (as at Waikaripi, Karen Wreck, Pumphouse and Kaingaroa). At all sites, growth estimates could have been affected by differential mortality of the smallest seed paua. Relationships anlong growth, mortality, Shell length (mm) and density were examined. There were no significant (m) correlations between annual B. ln~t~alseed slzes (20 Feb 91) C. At l0 months (11 Dec 91) 10 mortality and annual average growth incre- 40 ments (re = -0.472, ns), between growth in- 3o , crements and density (growth vs initial den- $ sity: r8 = 0.018, ns; growth vs final density: g 20 re = 0.602, ns), and between mortality and ini- tial density (r8= 0.130, ns). There was also no correlation between initial densities at seed- n ..... o . - .. - . - .- m- - -. ing and the densities on the final sampling - o j ib ik 2'0 2j 3b35 40 4j 50 o 5 10 15 20 25 30 35 40 45 YI 55 60 Shell length (mm) Shell length(mm) date in December 1991 (re = 0.108, ns). . D. Growlh E. Survival U! 35 , L0 Economic projections 30 . An analysis of the econonucs of enhance- ment at the Chatham Islands, based on the -2:E 5 15 . preliminary results of the experimental seed- m l9 ing (Table 4),must be treated as tentative be- - rj cause of the short time that paua were in 5 - -. S 0- -- 20 Feb/ 91 11Dec91 91 ll place and the variable times for which mor- X) Feb Dec 91 tality estimates were obtained at the different Date Date sites. The value of M, instantaneous mortality, Fig. 6. Hahotis Ins. Size-frequencies of abalone at Waikanpi (A) prior to seeding and (B, C) during the seeding experiment. (D) Average shell recorded at each site was used to assess pO- length (k 1 SE) of seeded paua through time. (E) Percentage survlval of tential survival for the first 3 yr after seeding the oriqinal cohort (f 95 % confidence interval) throuqh tune (Table 4). From Years 4 to 6, M = 0.2 was used, a value considered a reasonable and conserva- tial investment was based on the cost of seed from the tive estimate by Sainsbury (1982a) and others (Schiel & hatchery for each site, calculated at its future value of Breen 1991). The meat weight was based on average 5 % yr-' for the estimated 6 to 7 yr until paua are har- figures from the fishery (Schiel unpubl. data). The ini- vested. The internal rate of return (IRR) is the percent-

Table 3. Haliotis ~ns.Summary of enhancement results of abalone at Chatham Island sites. A: annual mortality rate; M instan- taneous mortality

Site Seeding Final mean % Survival Time at A M Average annual density (m-2) density (SE) (2CI) liberty (mo) increment (mm)

Port Hutt (Alabama) 19 7 10.6 (2.90) 53 8 (29.14) 23 0 276 0 323 22.6 Owenga 35 4 3 5 (0.80) 9 9 (4 44) 12 0901 2309 21.0 Port Hutt (2) 65 7 5.0 (1 24) 7 6 (3.78) 19 0 803 1 627 14.0 Whangamoe 29.0 7.0 (1.36) 24.2 (9.18) 19 0.592 0.897 23.8 Waikaripi 9 9 1.9 (0 52) 19.4 (10.35) 10 0.860 1.965 14.2 Karen Wreck 11.1 5.2 (1 12) 46 7 (20 09) 10 0.599 0 913 13.5 Pumphouse 17 0 5.0 (0.96) 29 1 (10 98) 10 0773 1.481 18.3 Kaingaroa 16.0 1.2 (0 60) 7.5 (7.35) 7 0 988 4.440 15.9 Mar. Ecol. Prog. Ser. 97: 167-181, 1993

Table 4. Haliotis iris. Economic projections for the financial returns of abalone seeded at 8 sites at Chatham Island. Projections calculated for each site as well as all 8 sites combined. M: instantaneous mortality from Table 3; IRR = internal rate of return (which needs to be 5% to break even). Assumphons were: (1) paua will take 6 yr to reach the minimum legal size of 125 mm; (2) survival for the first 3 yr (the juvenile period) is the figure recorded at each site during the experiment, then 80% afterwards (Sainsbury 1982a); (3) whole weight is 250 g paua-'; (4)meat recovery is 42 % of whole weight; (5) shell recovery is 30 % of whole weight; (6) the $ return is NZ$ 75 kg-' for meat and NZ$ 5 kg-' for shell; (7) the original value of the seed was calculated as $0.20 X number seeded at each site, the future value of this money was calculated at 5 % yr-'

Site M No. Meat Meat Shell Shell Total Future Profit/ IRR (%) of paua weight return weight return return value of loss (kg) (kg) investment

Port Hutt (Alabama) 0.323 Owenga 2.309 Port Hutt (2) 1.627 Whangamoe 0.897 Waikaripi 1.965 Karen Wreck 0.913 Pumphouse 1.481 Kaingaroa 4.440

l Projections based on costs and returns irom aii 8 sites combined: Total future value of investment in seed at the 8 sites = $20379 Total value of returns from 8 sites = $26349 Total profit on returns of the 8 sites = $5970 Internai raie oi reiurn for &heS siies = 3.6 70 I

age annual gain of future profits from the money in sand, and few juvenile paua were present. This invested at each site. source of mortality is consistent with what was seen in Only the 3 sites with relatively high survival rates finer-scale experiments within sites, where shifts in showed positive financial returns. Port Hutt (Alabama) rock, gravel, and sand affected the survival of juvenile showed the greatest potential returns and a 43.7 % paua (Sainsbury 1982a, Schiel 1992a). IRR. Whangamoe showed an IRR of 14.4 % and the It is not known to what degree movement of seed Karen Wreck site had an IRR of 9.3 %. The other 5 sites paua within and away from seeding sites affected esti- had negative returns. If the data for all 8 sites are com- mates of survival. Predominantly rocky sites such as bined (Table 4) there is an overall positive financial those at Port Hutt, Whangamoe, and along the Durham return of the value of paua over the future value of the coast had a large amount of adjacent habitat suitable invested money, and an IRR of 9.6 %. for juvenile paua. There is some evidence that juvenile paua can move hundreds of meters over a period of a few months (Pirker 1992) and other studies have DISCUSSION shown that abalone can move considerable distances (Newman 1966, Poore 197213, Shepherd 1986), so it is The major feature to emerge from this enhancement Likely that movement away from sites has occurred. study was the considerable variation in survival from The 2 sites seeded early in the study and monitored site to site. Mortality ranged from almost 100% after more than once showed that the lowest percentage re- less than 1 yr to as low as ca 50 % after almost 2 yr. The covery occurred at the first sampling. This may be the major identifiable source of mortality was the move- result of high mortality in the mo immediately follow- ment of sand and the complete or partial burial of ju- ing seeding, but may also reflect the difficulty in locat- venile habitat within sites. The long stretches of sand ing small paua in their cryptic habitat. Mortality rates on both the east and west coasts of Chatham Island were low after 1 yr. At one site, Port Hutt (Alabama), it render the sites near them vulnerable to this source of appeared that survival was greater at 2 yr than at 1 yr. mortality. Further evidence for the impact of sand on This may be the result of juveniles emerging from the habitats was seen at a site south of the Owenga seed- very heterogeneous rocky bottom and becoming less ing site. This site was surveyed in January 1990 as a cryptic as they get larger. An earlier experiment also potential seeding area. It had hundreds of m2 of juve- showed greater recovery after the first year (Schiel nile habitat and large numbers of juvenile paua. By unpubl. data). In this experiment, 2100 individually May 1990 the site had been almost completely buried marked paua were seeded, of which 18.7 % were re- Schiel Enhancement of abalone populations 177

captured after 1 yr and 24.3 % after 3 yr. It will require from a few studies. Shepherd (1987) found Mvalues of further monitoring to determine whether the estimates 3.5 to 4.6 for Haliotis laevigata aged 0 to 6 mo and of of survival at the other sites at Chatham Island follow a 0.2 to 1.1 for those aged 6 mo to 2.5 yr (quoted in similar trend. The large variance associated with later Shepherd & Breen 1992). This study showed decreas- estimates of survival indicates the characteristically ing mortality with increasing age. McShane (1991) aggregative behaviour of larger juveniles. found that survival of recruits (<30 mm) of H. rubra in Although most paua appeared to be healthy and Victoria, Australia, was as low as 0.3 % after 5 mo (M= active when they were seeded, it is likely that there 5.8), but that there was considerable variation among was some mortality associated with the seeding populations and among different years in the same process and the initial period when paua became population. His data also indicated that survival may established in natural habitats. In small-scale experi- be density-dependent. In my seeding trials, no density- ments, hatchery paua survival was ca 25 % lower than dependence was detected. The only natural mortality that of natural paua after 5 wk (Schiel 1992a). data available for H, iris showed that Mvalues fell into Laboratory-based experiments on Haliotis rufescens in the range of 0.02 to 0.25 for abalone >2 yr old California showed similar results, with hatchery-raised (Sainsbury 1982a). The results and implications from juvenile abalone being initially more susceptible to these studies are that initial mortality rates are quite predation and showing different movement patterns high for juvenile abalone of all species and that M compared to natural juveniles (Schiel & Welden 1987). decreases as abalone grow to larger sizes. These differences decreased after several days. Some data exist on the survival rates WREN WRECK SITE of small hatchery-raised abalone trans- planted to natural habitats, so comparisons ,6 A. Pre-seeding sizestructure can be made to this enhancement study. Tegner & Butler (1985) recorded M values of 0.4 to 0.5 for seeded H. rufescens in California (see also Shepherd & Breen 1992). The seed size, however, ranged from 45 to 71 mm, considerably larger than that used at Chatham Island. Saito (1984) recorded annual survival rates ranging from 0 2040608010012Q Shell length(mm) 6 to 20 % (M: 1.6 to 2.8) for H. discus hannai of 16 to 24 mm seeded onto natural reefs. B. Initial seed sizes (22 Feb 91) C. At 10 months(1 1 Dec91)

The best annual survival rate of 30 % was 40~ 1 obtained with seed of 22 to 38 mm. In another study with H. dscus hannai of 10 to 1/1fm 1: lllm,l 20 mm seed size, the Mvalues ranged from z 0.6 to 0.7 (Saito 1979, quoted in Shepherd & 10 , 10 Breen 1992).Some of the Mvalues recorded 5: at Chatham Island fell within the ranges of 0 these other studies. However, there were 0 5 10 15 ML.... 25 30 35 40 45 50 0 5 10 15 M 25 30 35 40 45 50 differences in how survival rates were esti- Shell length (mm) Shell length (mm) mated. Some studies apparently used re- D. Growth 30 capture rates of seed abalone (Saito 1984), V) 1 while others estimated survival using a combination of recapture of live abalone, counts of shells from dead abalone collected from seeding sites, and assumptions about the movement and survival of unseen abalone (Tegner & Butler 1985, 1989).

Because of the patchy and cryptic nature of 22 Feb 91 l1Dec91 22 Feb91 l1h91 small abalone, it is always difficult to esti- Date Date mate survival (Breen 1992, Shepherd & Fig. 7. Haliotis Size-frequencies of abalone at Karen Wreck site Breen 1992). iris. (A) prior to seeding and (B, C) during the seeding experiment. (D) Aver- Estimated instantaneous rates of natural age shell length (*l SE) of seeded paua through time. (E) Percentage mortality for wild abalone are available survival of the original cohort (f95 % confidence interval) through time Mar. Ecol. Prog. Ser. 97: 167-181, 1993

PUMPHOUSE SITE dictability in the size of recaptured paua, it may be necessary to minimize the range of ,, p.Pre-seeding sizestructure sizes of seed paua by rigorous size-grading in the hatchery. The problems of assessing growth and mortality highlight the cooperation that is necessary between the hatchery and field programs in enhancement work. Having seed abalone of a narrow size-range helps both in assessing growth and in identifying 0 10 20 30 40 M W 70 80 W 100 Shell length(mm) seed in future field surveys. Good hatchery techniques are essential to successful seed- B. Initial seed sizes(22 Feb 91) C. At 10 months (11 Dec 91) ing of abalone. Abalone that have been 35 r l4 1 stressed in the hatchery or are obviously stressed upon placing into natural habitats have little chance of survival (Tegner & Butler 1985, 1989, Schiel1992a). The greatest stresses to abalone occur in packing and transport (Ino 1966, Whang & Chung 1977). It was helpful to place paua into fast-flowing 0 5 10 15 20 25 30 35 40 45 50 55 M) seawater as soon as possible after transport. Shell length (mm) Shell length (mm) This removed faecal material from transport D.Growth - E. Suwivai tubes and ailowed paua to move around in 2 35 0; \O their tubes and feed on the algae inside. $ 30 Little mortality occurred after paua were ::l ::l acclimatised in this fashion. Further hatchery g cooperation is needed to mark or cohort-tag 5 15 C 40 . abalone in the large numbers necessary for -Q) 10 -Q) 5 \- commercial seeding. The ability to identify - &L paua as being of hatchery origin is essential 6 o 8 22 F& 91 11h91 z F& 91 11h91 in assessing the success of enhancement Date Date programmes. One effective means of tagging Fig. 8. Haliotis iris. Size-frequencies of abalone at Pumphouse site paua is feed them sequentiau~in (A) prior to seeding and (B. C) during the seeding experiment. the hatchery on different foods. By switching (D) Average shell length (+l SE) of seeded paua through time. the algae from Gracilana sordjda to Macro- (E)Percentage survival of the original cohort (+ 95 % confidence inter- cystis pyrifera every few weeks, for example, val) through time alternating bands of dark and bright green shell can be laid down. This pattern can be It is difficult to compare growth rates of seed Haliotis seen for several years and offers a way to tag hatchery ins among sites at Chatham Island because of the vary- paua cheaply and in large numbers. ing seasons that the seed were in place. Both Poore Another potential problem with hatchery seed is a (1972~)and Sainsbury (1982a, b) recorded average restricted genetic stock from using only a few parent annual growth increments of ca 20 to 25 mm for paua animals for spawning. Smith & Conroy (1992) recom- <20 mm in length. Pirker (1992) found that annual mend that seed should originate from at least 5 males growth of juveniles fell within this range but that there and 5 females to retain 95 % genetic variation of the were significant differences among sites, and between parent stock. The population consequences of reduced years within a site. At Chatham Island, annualized av- genetic variation of seeded abalone in terms of growth, erage growth rates ranged between 13.5 and 24.3mm. survival, and future breed~ngare unknown. Because of These estimates, however, reflect not only variations these uncertainties the parent stock should be as among sites but also the different seasons for which diverse as possible. paua grew at each site and possibly the effects of dif- In the case of enhancing commercially fished popu- ferential mortality of smaller seed. Some of the prob- lation~,the major reason for attempting enhancement lems of estimating growth rates would be obviated by is an economic return. The financial projections from having all seed of about the same initial size. To reduce these trials are uncertain because of the relatively the costs of raising seed and achieve greater pre- short periods for which growth and survival were esti- Schiel: Enhancement of abalone populations 179

mated at most sites. They were necessary, however, because of the long-term nature A. Pre-seedingslze structure of enhancement programmes and the need 10 to have an indication of whether future work is worth pursuing. One site in particu- lar, Port Hutt (Alabama), offers the best chances for a financial return on the initial investment. Other sites may also prove to be financially viable if future recapture rates are high. Some sites, particularly the sandy ones at Owenga and Kaingaroa, are inap- Shell length (mm) propriate for enhancement and have little chance of achieving adequate survival for B. Initial seed sizes(27 May 91) C. At 7 months (10 Dec 91) 200 - an economic return. I l The major conclusion from this study is that the variability in growth and survival can be reduced by choosing particular habi- tats and sites for enhancement. Selection of sites is difficult because areas where adults occur do not necessarily have juvenile paua or juvenile habitat. Although several stud- o 5 1015 20 25 30 35 40 45 50 o 5 10 15 20 25 30 35 40 45 50 Shell length(mm) Shell length(mm) ies have shown that juvenile abalone are found in rock crevices and beneath boul- . D. Growth f E. Survival ders (Poore 1972c, Shepherd & Turner 1985, 25 McShane et al. 1990, McShane & Smith $ 20 l 1991), it is clear from this study that the E habitat for seeding must be fairly extensive -a! and away from areas where sand move- l0 ment can destroy it. Movement of juveniles f 5 / V) 20. \ from cryptic habitats to open reef occurs 8 o.--. -- when paua reach 70 to 90 mm. The process 27 May91 10Dec 91 27 May91 10Dec 91 of dispersing to rocky reefs slong the coast Date Date may take several years (Poore 1972~). The results that strategies may Fig. 9. Haliotjs jns. Size-frequencies of abalone at Kaingaroa (A)pnor to seeding and (B, C) dunng the seeding experiment. (D) Average shell lead to One is to length (*l SE) of seeded paua through time. (E) Percentage survival of place seed in relatively small numbers in the original cohort (? 95 % confidence interval) through time numerous patches of habitat as they are encountered. This takes more effort in trans- porting seed paua to many sites, and presents risks in Transplantation of abalone is feasible, but has not stressing them. However, it spreads the risk of habitat- been done in New Zealand. A study on Haliotis related mortality over many different sites. Another kamtschatkana in Canada found that relaying stunted strategy is to locate as many larger sites as possible and individuals to favourable sites resulted in survival esti- use these as nurseries for juvenile paua. When these mates similar to the control site but increased growth grow to 70 to 90 mm they can be transplanted to other relative to the control site (Emmett & Jamieson 1989). sites for adult growth. This would take account of the This option requires further study in New Zealand and, pronounced differences in the habitat requirements of if viable, may offer the best hope of successfully seed- juvenile and adult paua. Juveniles feed on small algae ing paua in large numbers to fishing areas. and diatoms and live cryptically, while adults feed on It seems clear that there is no direct correlation drift and attached algae on open rock surfaces (Tun- between juvenile and adult numbers among sites, a bridge 1967, Poore 1972a, Garland et al. 1985). The result found for other species (Shepherd et al. 1992a). fishery for paua occurs in many places that seem to be Settlement requires particular chemical cues (Morse et far removed from juvenile habitat. By using nursery al. 1979), larval drift may be associated with topo- areas and transplantation, many of these good fishing graphic features of the coast (McShane et al. 1990), sites can be replenished without relying on the chance and recruitment is highly variable among sites and events of natural movement of paua into these areas. years (Sainsbury 1982a, McShane & Smith 1991). In 180 Mar. Ecol. Prog. Ser. 97: 167-181, 1993

sites where juvenlle habitat is abundant, there may be Japan. Nippon Suisan Shigen Hogo Kyokai, Suisan differences in available food (Poore 1972a, b), preda- Zoyoshoku Soshu 11. 1-104 (translated from Japanese by tion pressure, particularly by mobile reef fishes (Choat Trans. Ser. Fish. Res. Bd Can. 1078, 1968) Jones, G. P. (1990) The importance of recruitment to the & Ayling 1987), and exposure. The decoupling of num- dynamlcs of a coral reef fish population. Ecology 71: bers in the different life stages of paua means the iden- 1691-1695 tification of within-site sources of mortality is essential Keough, M. J. (1988). Benthic populations. is recruitment in understanding the population structure and dynam- hitlng or just fashionable? Proc. 6th int, coral Reef Symp. 1. 141-148 ics of abalone populations. This study has shown that McShane, P. (1991). Density-dependent mortality of recruits shifts in habitat, especially burial by sand, greatly of the abalone Haliotis rubra (: ). Mar. affects survival of juvenile paua. Therefore, by care- Biol. 110: 385-389 fully choosing sites, the risk of mortality due to habitat McShane, P, Black, K. P., Smith, M. G. (1990). Recruitment shifts can be greatly reduced. processes in Haliotis mbra (Mollusca: Gastropoda) and regional hydrodynamcs in south-eastern Australia imply The financial returns for abalone are increasingly localised dispersal of larvae. J. exp. mar. Biol. Ecol. 124: favourable to seeding of hatchery juveniles. For exam- 175-203 ple, the price of paua meat has increased by over McShane, P,, Smith, M. G. (1991). Recruitment variation In NZ$30 kg-' during the past 12 mo. The commercial sympatric populations of Haliotis rubra Leach (Gastropoda: Haliotidae). Mar. Ecol. Prog. Ser. 73: success of seeding will depend largely on the continu- 203-210 ation of this trend and the reduction of seed mortality Morse, D. E., Hooker, N., Duncan, H., Jensen, L. (1979). to that of the best sites in this study. It shouid aiso be Gamma-aminobii:yi5c acid, 2 zeuro!ransmitter, induces noted that, as in any wild fishery, enhancement with planktonic abalone larvae to settle and begin metamor- hatchery-raised juveniles is not a substitute for good phosis. Science 204: 407 Newman, G. G. (1966). Movements of the South African fisheries management. abalone Haliotis midae. Invest1 Rep. Div. Sea Fish. S. Afr. 56: 1-19 Pirker, J. G. (1992). Growth, shell-ring deposition, and mor- Acknowledgements. I thank the staff of MAF Fisheries and tality of paua (Haliotis iris Martyn) in the Kaikoura region. Salmon-Smith-Blolab for use of facihties. I thank Steve M.Sc thesis, Univers~tyof Canterbury, Christchurch Mercer, in particular, for his invaluable field assistance; John Poore, G. C. B. (1972a). Ecology of New Zealand , Schellevis for helping to set up the contract and Dr John HaLiotis species (Mollusca: Gastropoda) 1. Feeding. McKoy for his interest and insbtutional support; Graeme N.Z. J. mar. Freshwat. Res. 6. 11-22 Moss, John Illingworth for hatchery support; and paua fishers Poore, G. C. B. (1972b). Ecology of New Zealand abalones, Roger and Nicki Beattie. Peter Broad, Barry Lanauze, Andy Haliotis species (Mollusca: Gastropoda) 2. Seasonal and Hough, John Bates, and Eddie Reriti. Thanks to J. van Berkel diurnal movement. N.Z. J. mar. Freshwat. Res. 6: 246-258 for graphics, Drs P. Breen, P. McShane. L. Tong. M. Foster, Poore, G. C. B. (1972~).Ecology of New Zealand abalones, J. Pearse, and an anonymous referee for helpful criticism. Haliotis species (MoUusca: Gastropoda) 3. Growth. N.Z. J. mar. Freshwat. Res. 6: 534-559 Sainsbury, K. J. (1982a). Population dynamics and fishery management of the paua, Haliotis iris. 1. Population struc- LITERATURE CITED ture, growth, reproduction, and mortality. N.Z. J. mar. Freshwat. Res. 16: 147-161 Breen, P. A. (1992). A review of models used for stock assess- Sainsbury, K. J. (1982b). Population dynamics and fishery ment in abalone fisheries. In: Shepherd, S. A., Tegner, management of the paua, HaLiotis iris. 11. Dynamics and M. J , Guzman del Proo, S. A. (eds.) Abalone of the world; management as examined using a size class model. biology, fisheries and culture. Blackwells, Oxford, N.Z. J. mar. Freshwat. Res. 16: 163-173 p. 253-275 Saito, K. (1979). Studies on propagation of Ezo abalone, Choat, J H.,Ayhng, A. M. (1987). The relationship between Haliotis discus hannai Ino. I. Analysis of the relationship habitat structure and fish faunas on New Zealand reefs. between transplantation and catch. Bull. Jap. Soc. Sci. J. exp. mar. Biol. Ecol. 110: 257-284 Fish. 45: 695-704 Connell, J. J. (1985). The consequences of variation in initial Saito, K. (1984). Ocean ranching of abalones and scallops in settlement vs. post-settlement mortality in rocky intertidal northern Japan. Aquaculture 39: 361-373 communities. J. exp. mar. Biol. Ecol. 93: 11-45 Schiel, D. R. (1989). Species enhancement in the natural envi- Emmett, B., Jamieson, G. S. (1989). An experimental trans- ronment. Proceedings of AQUANZ'88 conference in plant of prerecrult abalone in Barkley Sound, Bntlsh Wellington, N.Z. (available from MAF Fisheries, PO Box Columbia. Fish. Bull. U.S. 87: 95-104 297, Wehgton, New Zealand) Garland, C. D.. Cooke, S. L., McMeekin, T A. (1985). Schiel, D. R. (1992a). The enhancement of paua (abalone) Ingestion of the bacteria on the cuticle of crustose (non- populations In New Zealand. In: Shepherd, S A., Tegner, articulated) coralline algae by post-larval and juvenile M. J.. Guzman del Proo, S. A. (eds.) Abalone of the world; abalone (Haliotis ruber Leach) from Tasmanian waters. biology, fisheries and culture. Blackwells, Oxford, J. exp. mar. Biol. Ecol. 91: 137-149 p. 474-484 Harrison, A. J. (1986). Gastropod fishenes of the Pacific with Sci-uel. D. R. (1992b). The paua (abalone) fishery of New particular reference to Australian abalone. Can. Spec. Zealand. In: Shepherd, S. A., Tegner, M. J., Guzman del Publ. Fish. Aquat. SCI.92: 14-22 Proo, S. A. (eds.) Abalone of the world; biology, fisheries Ino, T. (1966). The abalone sclence and its propagation in and culture. Blackwells, Oxford, p. 427-437 Schiel: Enhancement of abalone populations 181

Schiel, D. R., Breen, P. A. (1991). Population structure, ageing Smith, P. J., Conroy, A. M. (1992). Loss of genetic variation in and fishing mortality of the New Zealand abalone Hahotis hatchery produced abalone, Haliotis iris. N.Z. J. mar. iris. Fish. Bull. U.S. 89: 681-691 Freshwat Res 26 81-85 Schiel, D. R., Welden, B. A. (1987). Responses to predators of Tegner, M. J , Butler, R A (1985). The survival and mortality cultured and wild red abalone, Haliotis rufescens, in labo- of seeded and natlve red abalones, Haliotis rufescens, on ratory experiments. Aquaculture 60: 173-188 the Palos Verdes Peninsula. Calif. Fish Game 71 150-163 Shepherd, S. A. (1986).Movement of the Southern Australian Tegner, M. J., Butler, R.A (1989) Abalone seeding. In: Hahn, abalone HaLiotis laeviyata in relation to crevice abun- K. 0. (ed.) Handbook of culture of abalone and other dance. Aust. J. Ecol. 11: 295-302 manne gastropods. CRC Press, Boca Raton, p. 157-182 Shepherd, S. A. (1987). Aspects of the biology of the abalone Thayer, G. W. (ed.) (1992). Restor~ngthe nation's marine envi- Haliotis laevigata and HaLiotjs scalarjs. Ph.D. thesis, ronment. Maryland Sea Grant College, University of Deakin University Maryland, College Park, Shepherd, S. A., Breen, P. A. (1992). Mortality in abalone: its Tong, L. J., Moss. G A. (1992). The New Zealand culture estimation, variabihty and causes. In: Shepherd, S. A., system for abalone In: Shepherd, S. A., Tegner, M. J., Tegner, M. J., Guzman del Proo, S. A. (eds.) Abalone of the Guzman del Proo, S. A. (eds.) Abalone of the world; biol- world; biology, fisheries and culture. Blackwell, Oxford, ogy. fisheries and culture. Blackwells, Oxford, p. 583-591 p. 276-304 Tunbridge. B. R. (1967). Feeding habits of paua (Haliotis iris Shepherd, S. A., Lowe, D., Partington, D. (1992a). Studies on Martyn). Fish. tech. Rep. N.Z. mar. Dep. 20: 1-18 southern Australian abalone ( Haliotis) XIII: larval Underwood. A. J.. Denley. E. J. (1984). Paradigms, explana- dispersal and recruitment. J. exp. mar. Biol. Ecol. 164: tions and generalizations in models for the structure of 247-260 intertidal communities on rocky shores. In: Strong, D., Shepherd, S. A.. Tegner, M. J., Guzman del Proo, S. A. (eds.) Simberloff. D., Abele, L. G.. Thistle. A. B. (eds.) Ecological (1992b). Abalone of the world; biology, fisheries and cul- communities: conceptual issues and the evidence. ture. Blackwell, Oxford Princeton University Press. Princeton. p. 151-180 Shepherd, S. A., Turner, J. A. (1985). Studies on southern Whang, H. C., Chung, K. 0. (1977). Study on relationship Australian abalone (genus Haliotis). VI. Habitat prefer- between exposure time and mortality of young abalone, ence and abundance and predators of juvedes. J. exp. Haliotis discus (Reeve). Bull. Fish. Res. Devel. Agency, mar. Biol. Ecol. 93: 285-298 Busan 18: 123-129

This article was submitted to the editor Manuscript first received: November 9, 1992 Revised version accepted: May 10, 1993